Digital Connectivity and GovTech Maturity Across 162 Economies:Development-Adjusted Analysis and Interpretable Machine Learning for Public Information Systems

Abstract

Digital government operates within a connectivity environment that can facilitate or hinder the use of public information systems. This study examines the cross-national association between digital connectivity and GovTech maturity while explicitly accounting for income and regional structure. A cross-sectional secondary-data design integrates the International Telecommunication Union ICT Development Index (IDI) 2025, the World Bank GovTech Maturity Index (GTMI) 2025, and United Nations E-Government Survey 2024 indicators. After ISO-3 matching, 162 economies had valid GTMI observations. Pearson correlations assessed linear relationships and Spearman correlations assessed monotonic rank relationships. HC3-robust regression controlled for World Bank income group and ITU region. Ridge regression and random forest were evaluated with 20 repeated five-fold out-of-fold runs and a leave-one-ITU-region-out robustness check. IDI was associated with GTMI (r = 0.601, p < 0.001); after income and region adjustment, a 10-point higher IDI remained associated with a 0.108 higher GTMI (p < 0.001), with an adjusted partial correlation of r = 0.404. Cross-source convergence was strong for public-service delivery versus the UN Online Service Index (r = 0.813) and digital citizen engagement versus the UN E-Participation Index (r = 0.819). Random forest achieved modestly higher repeated out-of-fold R² than ridge (0.367 vs. 0.294; mean ΔR² = 0.073, 95% CI 0.057–0.090), and region-grouped validation yielded R² = 0.334. Held-out permutation analysis and grouped sensitivity analysis identified affordability as the strongest model-level signal, although correlated predictors limit individual attribution. The findings show that connectivity remains meaningfully associated with GovTech maturity after broad development controls, but accounts for only part of cross-national variation. The results support affordability-aware, mobile-first, low-bandwidth, interoperable, and citizen-centered public information-system design.